-
Literature and corpus. The review added the sources the names question depends on. Kelley's 2018 thesis shows that the 'herded humans' accounts behind Englund's names are under 1% of the corpus [@kelley2018gender]. Bartash 2018 covers how the earliest administrators recorded people by age and gender [@bartash2018age]. Dahl's 2025 synthesis dates the first phonetic name spellings to Archaic Ur [@dahl2025early]. Biggs and Pomponio give the Semitic and Fara name material [@biggs1967semitic] [@pomponio1984fara]. Andersson 2012 is the reference for Early Dynastic Sumerian name elements [@andersson2012kingship], and Lecompte 2013 adds Archaic Ur tablets [@lecompte2013archaic]. Born and Kelley's warning that within-case sign order in transliteration is partly the editors' choice shaped the method [@born2021tribute].
-
Corpus. The Uruk IV–III transliterations came from the Nino-cunei/uruk copy of CDLI (6,398 tablets). Every Early Dynastic I–II, IIIa and IIIb inscription on CDLI (6,996) was exported, and the 394 Archaic Ur tablets transliterated in sign names were stored as a dataset. In the CDLI convention the editors mark personal names with
( )a, so the name set is theirs, essentially Englund's, not a heuristic: 602 entries, 540 with a legible sign, on 41 tablets. Oracc's tagged name corpora could not be reached over TLS from the workspace, which blocked the planned comparison with Early Dynastic name lists. -
Methods assessment. A direct comparison with Early Dynastic names was not possible this session. Positional statistics were demoted to exploratory, because the order is partly editorial. The confirmatory test chosen compares names with the corpus's own non-name entries, using Englund's list of Sumerian name elements fixed in advance.
-
Main result. 12.4% of names contain one of Englund's elements, against a null median of 9.8% (95%: 7.4–12.4%). The odds ratio is 1.35 (1.09–1.60) against the corpus and 0.97 (0.70–1.38) against the same tablets. Under the pre-declared rule this is no preference. The deficit Englund's argument implies is not there.
The archaic names carry Englund's Sumerian elements at about the corpus rate, not below it
Share of 540 editor-marked names containing at least one element, against 10,000 random pseudo-names of the same lengths; CDLI Uruk III–IV transliterations
-
Follow-ups. Adding EN gives an enriched set against the corpus (26.5% against 13.7–20.0%), but the excess lies entirely in the unprovenanced Schøyen and Brussels accounts. In the excavated archives the share is 16.1%, within a null of 12.1–25.0%. In the editors' order, EN stands first in 54 of 80 multi-sign names that contain it (shuffle range 25–42), a shape like the Sumerian en-X names, pending collation.
-
What it changes. The names no longer count as measured evidence against Sumerian. They don't count for it either. The central question now rests on phonetic spellings and lexical-list continuity, and on a calibrated name benchmark once the Oracc data can be reached.